02 · China Bond Market in Practice
China's bond market is the world's second largest, yet over 90% of trading happens in the interbank market — the institutional playground. Retail investors' real arena is the exchange market, and there's plenty to play with: government bond reverse repos, savings bonds, bond funds, convertible bonds.
This chapter dives into how China's bond market actually works: get the structure straight first, then dissect every retail-accessible tool in detail, and finally see clearly the institutions' carry game and the risk truth behind "breaking implicit guarantees."
I. Market Structure: Interbank vs Exchange
| Market | Participants | Share of volume | Can individuals participate? |
|---|---|---|---|
| Interbank market | Banks, insurers, funds, brokers and other institutions | ~90%+ | Not directly (only indirectly via funds, etc.) |
| Exchange market | Institutions + qualified individual investors | Under ~10% | Can trade Treasuries, local government bonds, corporate bonds, convertible bonds |
Outstanding Stock Composition (common-knowledge figures; defer to latest data)
| Instrument | Issuer | Approximate share | Risk |
|---|---|---|---|
| Government bonds (Treasuries) | Ministry of Finance | Around 20% of outstanding | No default risk in local currency; benchmark rates |
| Local government bonds | Local fiscal authorities | Largest single instrument class in recent years (~40%) | Backed by local fiscal credit; low risk |
| Financial/policy bank bonds | Policy banks, commercial banks, etc. | Around 20% of outstanding | Low risk; mainstream institutional allocation |
| Credit bonds (enterprise/corporate bonds, MTNs, commercial paper) | Corporates | Around 20% of outstanding | Depends on issuer quality; highly dispersed |
📖 Why Retail Lives on the Exchange
The interbank market is a wholesale market between institutions — minimum trade sizes, account systems, and quoting habits are all built for them. The exchange market is the counter prepared for retail investors. So "the retail China bond market" ≈ exchange-traded Treasuries/corporate bonds/convertibles + over-the-counter savings bonds and funds.
II. Government Bond Reverse Repos: The Retail "Cash-Like" Tool
A government bond reverse repo = you lend money, the counterparty pledges government bonds, principal plus interest is repaid at maturity. In essence it's "overnight/short-term lending collateralized by Treasuries" — credit risk is extremely low (collateral is government paper).
Codes and Thresholds (subject to the latest regulations/policy)
| Market | Instrument | Code | Entry threshold (reference) |
|---|---|---|---|
| Shanghai | GC001 (1 day) and the GC series | 204001 etc. | From 100,000 RMB |
| Shenzhen | R-001 (1 day) and the R series | 131810 etc. | From 1,000 RMB |
- You act by choosing "sell"/"lend securities" (you are lending funds); just enter the code and annualized rate, and funds return automatically at maturity.
- Funds usable T+1, withdrawable T+2 (subject to each broker's latest rules).
When Are Rates High?
- Month-end, quarter-end, year-end: assessment points for banks/institutions tighten funding; reverse repo rates spike seasonally.
- Before long holidays (Spring Festival / National Day / May Day): demand for funds peaks; historically 3%–10% or higher annualized was common (subject to the latest conditions).
- Normal times: usually only 1%–2%, fluctuating with the pace of central bank liquidity injections.
- Pattern: the rate center is set by funding tightness, and it reliably rises around quarter/year/holiday crossings.
How Returns Are Calculated (numeric example, teaching approximation)
Return ≈ lent amount × annualized rate × actual days funds are occupied ÷ 365- Buy GC001 one day before National Day at a 6% annualized rate, funds occupied 8 days (holidays count):
- On 100,000 RMB: ≈ 100,000 × 6% × 8 ÷ 365 ≈ 131.5 RMB
- The same 100,000 RMB on a normal day at 1.8% for 1 day: ≈ 100,000 × 1.8% × 1 ÷ 365 ≈ 4.9 RMB
- Net of fees (on the order of one hundred-thousandth), pre-holiday returns clearly beat same-period money funds and demand deposits.
Versus Money Funds / Deposits
| Tool | Yield elasticity | Liquidity | Suited for |
|---|---|---|---|
| Demand deposits | Low | Anytime | Everyday pocket cash |
| Money market funds | Stable (roughly the 7-day repo range) | T+1 | Daily cash management |
| Reverse repos | Low normally, spiking at key moments | Back at maturity | Month/quarter ends and pre-holiday "picks" |
💡 Practical Tip: Watch the Key Dates
Watch reverse repo rates 1–2 days before long holidays and on the last 1–2 trading days of a quarter; switch if they're clearly above money funds. Don't bother otherwise.
III. Savings Bonds: The Counter's "Risk-Free Deposit"
Savings bonds come in two types, purchasable only at bank counters/online banking, not tradable on secondary markets:
| Type | Certificate form | Rate features | Early redemption |
|---|---|---|---|
| Certificate-type | Paper certificate/electronic record; lump-sum principal + interest at maturity | Fixed at issuance | Tiered interest accrual after 6 months of holding, with a handling fee; early redemption forfeits some interest (subject to the latest rules) |
| Electronic-type | Electronic account; interest paid annually | Fixed at issuance, interest arrives yearly | Early redemption deducts interest and fees, and usually only full redemption is supported (subject to the latest rules) |
- Threshold: from 100 RMB (subject to the latest rules); issued periodically through the year and often sells out instantly — popular tenors (3/5 years) require a scramble.
- Rate character: somewhat above same-term time deposits, less flexible than secondary-market Treasuries, but no mark-to-market risk (hold to maturity and the coupon is paid as contracted).
- Suited for: extremely conservative money certain not to move within 5 years.
💡 In an Easing Cycle, the Earlier You Lock, the Better
Savings bond coupons declining year after year is historical normal (they track easing cycles): "grabbing an issue means locking today's rate." In an easing cycle, earlier locking beats later.
IV. Bond Funds: A Risk-Return Hierarchy
| Type | Investment scope | Historical annualized range (reference) | Typical max drawdown (reference) | Risk |
|---|---|---|---|---|
| Short-term bond funds | Mostly bonds under 1 year | 2%–3% | Tiny (a few thousandths) | Very low |
| Pure bond funds (intermediate/long) | Bonds only, no equities | 3%–5% | 2%–4% | Low |
| Hybrid bond funds (secondary) | ≥80% bonds + ≤20% stocks/convertibles | 4%–7% | 5%–10% | Medium |
| Convertible bond funds | Mainly convertibles | Highly elastic: can double in bull markets, halve in bears | 15%–40% | High |
"Hidden Leverage" and NAV Volatility
- Bond funds commonly lever up via repo financing (public bond fund leverage cap ~140%, subject to the latest regulations/policy): the fund company borrows to buy more bonds, amplifying both returns and volatility — the NAV swings you see are post-leverage results.
- Bond funds can lose money too: in November 2022's domestic bond market correction, many pure bond funds drew down more than 1% in a month (seismic for low-risk products), mainly due to rapidly rising rates and a negative feedback loop of wealth-management redemptions. Lesson: bond funds ≠ "can't fall"; short-duration holds up better than intermediate/long, and redemption waves hurt more than the hikes themselves.
💡 Three Things to Check When Picking a Bond Fund
Is duration long (rate risk)? Is leverage high (amplification factor)? Is the convertible share large (a source of volatility)? — their combination determines NAV volatility.
💀 Iron Rule: Bond Funds Are Not "Can't-Fall" Assets
Bond funds can lose money — in November 2022's domestic bond correction, many pure bond funds drew down more than 1% in a month (seismic for low-risk products), mainly due to rapidly rising rates and the wealth-management redemption feedback loop. Lesson: bond funds ≠ "can't fall"; short duration resists better than intermediate/long, and redemption waves hurt more than hikes themselves. So treating a bond fund as a "deposit replacement" invites real NAV losses when hikes and redemption waves hit together.
V. Convertible Bonds Deep Dive
A convertible = bond + free option: within the agreed window it can be converted into stock at the conversion price. Bond floor below, equity upside above, T+0 trading, no daily price limits (circuit breakers added under newer Shanghai/Shenzhen rules; subject to the latest regulations/policy).
5.1 Bond Floor and Conversion Premium
| Concept | Meaning |
|---|---|
| Bond floor (straight-bond value) | Its value as an ordinary bond — coupon plus redemption price discounted; forms the price "floor" |
| Conversion value | Current stock price × 100 ÷ conversion price |
| Conversion premium | (Convert price ÷ conversion value − 1). High premium = weak equity character, sluggish upside follow-through |
- Premium ≈ 0 → the convert moves almost in lockstep with the stock (equity-like); high premium → rises slower than the stock (bond-like).
- The lower the price, the closer to the bond floor: near par of 100, downside is cushioned by the floor while upside tracks the stock — the "downside protection" story only truly holds here.
5.2 Forced Redemption Clause (The Issuer's "Death Sentence")
- If the underlying stock closes at ≥ 130% of the conversion price for a stretch of days (e.g., 15 of 30), the issuer may redeem at par plus accrued interest (terms per each convert's prospectus).
- After a forced redemption announcement, converts often fall from 130+ toward 100–105 — holders who neither convert nor sell get bought out at 100, vaporizing the premium.
- Practical meaning: never hold a convert nearing forced redemption as an "unrealized bull story."
💀 Iron Rule: Never Hold a Convert Near Forced Redemption
Never treat a convert nearing forced redemption as an "unrealized bull story." After the announcement, converts often fall from 130+ toward 100–105 — holders who don't convert or sell get bought back at 100, vaporizing the premium. So the biggest risk in convertibles isn't the underlying stock but clause games: read each issue's prospectus terms, and never mistake "floor" for "principal guarantee."
5.3 The Double-Low Strategy (Entry-Level Parameters)
- Double-low = low convert price + low conversion premium; classic screen: price < 110 (or 115), premium < 30% (parameters vary by strategy and market conditions; subject to the latest data), then rank and diversify across names.
- Logic: low price → near the bond floor, limited drawdown; low premium → follows the stock as soon as it rallies.
- Caveat: the double-low strategy buys patience for volatility — if the stock doesn't rise during your holding period, the convert drifts sideways or bleeds slowly (time cost). It is not a "guaranteed return" strategy.
5.4 The "Downward Revision Game": Relatively Higher-Certainty Opportunities
- Downward revision clause: when the underlying falls sharply, the issuer may lower the conversion price (requires shareholder meeting approval); conversion value jumps, and the convert usually gaps up.
- Why would issuers revise down? Because they don't want to repay the debt — pushing the stock toward conversion turns debt into equity.
- Game-theory essentials:
- High yield to maturity + major shareholders motivated to convert (e.g., large convert share outstanding, company short of cash) → high probability of revision;
- But revision is a "right," not an "obligation"; announcements are uncertain, and shareholder meetings have genuinely voted revisions down.
- Discipline: enter only once a revision lands and the price is reasonable; never go heavy on rumors of a "possible revision."
⚠️ Clause Games Run Deep — Read Each Issue's Terms
Convertibles are retail investors' advanced "attack-and-defense" tool in the bond market, but clause games run deep: read each issue's prospectus terms (conversion price, put, forced redemption, revision thresholds all differ), subject to the latest announcements.
VI. The Institutional Games of the Bond Market
Carry Trade: Borrow Short, Buy Long
Logic: borrow short-term funds at low rates (repos), buy longer-dated bonds yielding more
Return = long-end yield − short-end funding cost = term spread (carry)
Risk = when rates rise, falling long-bond prices eat the carry- When the curve is steep (low short end, high long end), carry is fat and institutions lever up eagerly; when flat/inverted, carry compresses to zero or negative and institutions are pushed into directional rate trades.
- Why retail can't copy this: institutions lever via repo financing (leverage of several to a dozen-plus times), amplifying both returns and volatility; retail has no funding access (individual repo thresholds are extremely high) and can't withstand the forced liquidation pressure that levers positions face when rates rise.
Leverage: A Double-Edged Sword
- In bond bull markets, institutions feast on "leverage + coupon"; once rates rebound, deleveraging accelerates the decline (the November 2022 "redemption → selling → more redemption" spiral being the classic case).
💡 Why Understanding Institutional Games Matters
Reading institutional games isn't about imitation — it's about finding low prices at the moments institutions are forced to liquidate (panic selling) and avoiding crowded, richly valued trades.
✅ Conclusion: Find Low Prices When Institutions Are Forced Out; Avoid Crowded Trades
Reading institutional games isn't imitation — it's finding low prices at forced-liquidation moments (panic selling) and avoiding crowding. In bond bull markets, institutions profit handsomely from "leverage + coupons"; once rates rebound, deleveraging accelerates declines (November 2022 being typical). Retail's correct stance isn't to follow the leverage but to wait for such "stampede" moments and pick up bargains.
VII. China Bond Market Risks: Breaking Implicit Guarantees
The Process of "Breaking Implicit Guarantees"
- 2014: the "11 Chaori Bond" default (China's first substantive credit bond default); 2015–2016: chain defaults like the "12 Shengda Bond" — the belief that "SOEs never default" began to crack.
- 2016: the "15 Tewoo" event — a centrally-affiliated issuer suspended payments and the market shook; afterward implicit-guarantee expectations were thoroughly broken and credit bond pricing shifted to genuine risk pricing.
- Private-enterprise defaults became routine (every year; specific cases subject to the latest market data), and the myth of "stable returns" from bond funds/wealth products dissolved.
The Divergence of Chengtou (LGFV) Bonds
| Historical phase | Characteristics |
|---|---|
| Past | Local government financing platform (chengtou) debt was treated as quasi-sovereign credit with strong implicit guarantees |
| Present (subject to the latest regulations/policy) | As debt-resolution policies advance, risk pricing diverges between high-debt regions and weaker platforms; "on-balance-sheet bond guarantees" coexist with "off-balance-sheet defaults" — public bonds and non-standard instruments are two different things |
| Core discipline | Chengtou bonds are not "risk-free"; any "XX will never default" narrative is not a reason to buy |
Three Cognitive Upgrades for Ordinary Investors
- "Bonds can lose money too": rate risk and credit risk show up in real NAVs (November 2022 being the most recent lesson for everyone).
- "High yields always come with an explanation": a bond yielding far above its rating cohort at the same maturity is being tagged with default risk — it isn't a free lunch.
- "Don't step outside your circle of competence": direct credit bond purchases require qualified-investor thresholds (historically ~1 million RMB; subject to the latest rules); retail's correct route is indirect participation via funds, leaving credit research to professionals.
VIII. Quick Reference: How Ordinary Investors Access Bonds
The earlier sections dissected each retail tool in turn; here is a "threshold × investor type" cheat sheet:
| Route | Threshold | Suits | Characteristics |
|---|---|---|---|
| Savings treasuries | From 100 yuan (per latest rules) | Conservative | Certificate/electronic; hold to maturity for principal + interest; not transferable |
| Exchange treasuries | ~1,000 yuan face value | Those with a securities account | Real-time trading; price moves with rates |
| Treasury reverse repo | Shenzhen from 1,000 yuan | Idle-cash holders | Nearly risk-free; big pre-holiday yield pops |
| Pure/short bond funds | From 10 yuan | Steady type | Low threshold, good liquidity, small swings |
| Convertible bonds | ~100 yuan face value per bond (one lot of 10 ≈ 1,000 yuan, per latest rules) | Advanced | Debt floor below, equity upside above; must learn the terms |
| Bond ETFs | On-exchange | Liquidity seekers | E.g., treasury ETFs, corporate bond ETFs (codes per latest) |
Suggested beginner path: treasury reverse repo (learn the mechanics) → savings treasuries (lock the risk-free return) → pure bond funds (delegate to professionals) → convertibles (only after advancing).
Risk Warning
⚠️ Risk Warning
China's bond market is no "sure win": rising rates drag down bond fund NAVs; credit bonds carry substantive default risk — after implicit guarantees broke, "capital guaranteed" no longer exists; convertibles bring clause risks (forced redemption, downward revision, investor puts) plus transmission from the underlying stock, and T+0 raises the bar for trading discipline; even Treasury reverse repos, though collateralized, carry a small probability of counterparty and liquidity risk. All codes, thresholds, rates, and drawdown figures here are teaching approximations — defer to the latest market conditions and the latest regulations/policy. This article is not investment advice.